List your company

GUI Agent Robustness Evaluation on Pop-up Perturbation 1.0 (Small)

97DSRCA
11.7233.865678.14Jul 13, 2025

Evaluation results

MethodLinks
CABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.07971003
CABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Default2025.079787.49.6
CABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0796.51003.5
CABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.079485.48.6
CABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.0793.51006.5
CABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0792.51007.5
CABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.078791.54.5
CABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0785.590.55
NDBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0764.381.417.1
NDBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.07617716
NDBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0759.576.517
NDBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.0758.879.921.1
DPOBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.07467428
DPOBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0745.57226.5
DPOBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.0743.774.430.7
DPOBackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0742.272.430.2
DABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0741.575.534
DABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.0741.578.537
DABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0741.260.319.1
DABackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.0737.764.827.1
DABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0732.573.541
DABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.0732.57542.5
NDBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.0725.164.339.2
DABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0721.674.452.8
DABackbone=LLaVA-v1.6-Vicuna-13B (LaSM L12-28, α=1.2), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.0721.674.452.8
NDBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.07216746
NDBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0720.665.845.2
DPOBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Highlight2025.0720.661.841.2
NDBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Default2025.0719.56747.5
DPOBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Overlay, Visual Prompting=Default2025.0718.165.847.7
DPOBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Highlight2025.0715.563.548
DPOBackbone=Qwen2-vl-7B (LaSM L7-18, α=1.1), Injection Type (IT)=Inductive, Visual Prompting=Default2025.07156550